A Real-World Example Of The Power Of AI In Agriculture

James Veale

During meetings with our partners in Singapore recently, the head of one practice reminded me that the world of agriculture and emerging technologies – such as the Internet of Things, machine learning, artificial intelligence, blockchain, drone imagery, and geo-information systems – are made for each other. He argued that with vast heactares of land, lake, or sea under farm management, it is impossible to get an accurate view of your agribusiness without multiple data inputs from (say) inexpensive digital technology devices.

He further argued that in the engineering sector (think car factory) and consumer products industries (think Coca-Cola bottling plant), pretty much the entire plant is instrumented and information-system controlled to enable the most efficient operation, and this rich data stream allows employees and managers to make the best business decisions to maximize plant efficiency. The agri equivalent would be to maximize yield.

To me, it is obvious that collecting many individual data points from your agri operation into your digital core aids analysis of problems, speeds up decision making, and leads to a genuine benefit for your enterprise. But what role can machine learning and AI play in the agri space?

Here is a great example from Global Fishing Watch. GFW uses data and open technologies to protect the oceans by showing, for free, exactly where all the trackable commercial fishing activity has happened since 2012 and is happening today in near-real time. The GFW data is open to the public, increasing transparency about fisheries worldwide; I encourage you to check it out.

To protect inshore waters from illegal fishing, Indonesia is the first nation to share all its small vessel VMS (vessel monitoring system) and large vessel AIS (automatic information system) tracking data in near real time. What is more, human operators have trained AI to recognize different vessel movement signatures from this captured data. The GFW system can recognize the difference between vessels that are merely passing through Indonesian inshore waters from those that are fishing. The system can even distinguish between different types of fishing activities – from trawling to line fishing to crab and lobster pots – with better than 95% accuracy.

Already, multimillion-dollar fines have been levied from the evidence gathered by this system, but the true power is its ability to look at the Big Data in near-real time and correctly identify the vessels’ activity, which can then be cross-referenced against fishing permits to see if the activity is legal or not.

Indonesia will ensure a sustainable fishing industry for its 350 million inhabitants for decades to come, but imagine what AI can do for your agri business. Time-consuming crop monitoring can be replaced with a range of inexpensive sensors, and this will substantially increase the precision and volume of data collected. With microscopic data collection, farmers will be able to produce diagnostics specific to individual plots or even single plants.

Likewise, human facial recognition technology (think iPhone 10) has been adapted to recognize individual cattle, which eliminates the stress of fitting cattle sensor devices, allowing easy monitoring of an entire heard with minimal interaction. This also enables individual monitoring of group behavior, early detection of lameness, and accurate recording of feeding habits.

Get more insight from research on How Sensors Will Redefine Business and Our World.


James Veale

About James Veale

James Veale is the Digital Leader for Agriculture industries at SAP Asia-Pacific and Japan. He leads the industry through value management, customer co-innovation, digital transformation, and business process performance improvement programs by developing road maps, reimagining business models, and reducing costs with digital technologies.